Free CLAUDE.md Generator
A CLAUDE.md generator turns a few facts about your repo (language, framework, commands, conventions, gotchas) into a ready-to-paste project memory file that Claude Code reads on every session. Drop the output at the root of your repo and Claude follows your rules from the next prompt onward, no extra setup needed.
Project
Commands
Conventions and gotchas
CLAUDE.md
55 lines · updates as you type
# Acme Web App
Customer-facing dashboard for the Acme platform. Next.js App Router on Vercel, Postgres via Prisma.
## Stack
- **Language:** TypeScript
- **Framework:** Next.js 15 (App Router)
- **Package manager:** pnpm
## Commands
```sh
# Dev
pnpm dev
# Build
pnpm build
# Test
pnpm test
# Lint / typecheck
pnpm lint && pnpm typecheck
```
## Coding conventions
- Use server components by default; mark client components with "use client"
- Tailwind classes only, no inline styles
- Prefer named exports over default exports for components
- Co-locate component tests in *.test.tsx next to the source file
## Gotchas
- Never run prisma migrate dev against production - use prisma migrate deploy in CI
- The /api/auth/* routes are auto-generated; do not edit them by hand
- Sentry tags are required on every server action - import { withSentry } from @/lib/sentry
## Glossary
Org: a customer tenant (one Org maps to one billing account)
Workspace: a project inside an Org (Orgs can have many Workspaces)
Run: a single execution of a workflow inside a Workspace
## Hard rules
- Never edit files in src/generated/** (auto-generated from the schema)
- Always run pnpm typecheck before claiming work is done
- Prefer fixing the root cause over adding try/catch suppression
---
_Generated with the free [CLAUDE.md Generator](https://www.tornic.dev/tools/claude-md-generator) by Tornic._
How to write a CLAUDE.md file
- Describe your project. Enter the project name, a one-line description, the primary language, and the framework or runtime you use.
- Add your commands. Paste the exact dev, build, test, and lint commands. Specific commands beat vague descriptions because Claude can run them verbatim.
- List conventions and gotchas. Note formatting rules, naming conventions, and the bugs or footguns Claude should avoid. Concrete examples beat abstract rules.
- Add a glossary and custom rules. Define internal terms, acronyms, or domain words. Add any hard rules ("never edit migrations", "always use the TaskQueue helper").
- Copy or download. Click Copy to grab the markdown, or Download to save it as CLAUDE.md. Drop the file at your repo root and Claude Code reads it on the next session.
Frequently Asked Questions
What is CLAUDE.md?
+
CLAUDE.md is a plain Markdown file Claude Code reads at the start of every session for project-specific instructions. It tells Claude how to run, build, and test your project, what conventions to follow, and what mistakes to avoid. Drop it in your repo root and Claude Code picks it up automatically.
Where do I put CLAUDE.md?
+
The repository root is the standard place. Claude Code also reads ~/.claude/CLAUDE.md for global personal instructions, and nested CLAUDE.md files in subdirectories for context that only applies inside that subtree. The repo-root version is what most teams need.
What should CLAUDE.md include?
+
The most useful sections are: a one-paragraph project overview, exact build / test / lint / dev commands, coding conventions (formatting, naming, imports), common gotchas and footguns, a project glossary for domain-specific terms, and any hard rules about what Claude should never do (e.g., never edit auto-generated files).
Does Claude Code automatically read CLAUDE.md?
+
Yes. Claude Code loads CLAUDE.md at session start with no configuration needed. The contents become part of the system context for every prompt in that session, so anything you write there shapes Claude's behavior throughout.
What is the difference between CLAUDE.md and AGENTS.md?
+
CLAUDE.md is Claude Code's native project-instructions file. AGENTS.md is a cross-tool convention used by some other coding agents (Cursor, OpenAI Codex CLI, and others). Many teams keep both files with the same content, or symlink one to the other so any agent gets the same guidance.
How long should CLAUDE.md be?
+
Aim for under 200 lines. CLAUDE.md is loaded into context on every turn, so every line costs tokens. Keep it scannable: bullets over paragraphs, exact commands over vague descriptions, and link out to deeper docs instead of pasting them inline.
Want Claude Code to follow the same plan every run?
CLAUDE.md sets the rules. Tornic enforces them. Turn your Claude Code, Codex, or Cursor CLI into a deterministic workflow engine that runs the same predictable steps every time, with no surprise API bills and no agent drift.
Try Tornic